Refined Characterization of the Expression and Stability of the SMN Gene Products

【摘要】  Spinal muscular atrophy (SMA) is characterized by degeneration of lower motor neurons and caused by mutations of the SMN1 gene. SMN1 is duplicated in a homologous gene called SMN2, which remains present in patients. SMN has an essential role in RNA metabolism, but its role in SMA pathogenesis remains unknown. Previous studies suggested that in neurons the protein lacking the C terminus (SMN7), the major product of the SMN2 gene, had a dominant-negative effect. We generated antibodies specific to SMNFL or SMN7. In transfected cells, the stability of the SMN7 protein was regulated in a cell-dependent manner. Importantly, whatever the human tissues examined, SMN7 protein was undetectable because of the instability of the protein, thus excluding a dominant effect of SMN7 in SMA. A similar decreased level of SMNFL was observed in brain and spinal cord samples from human SMA, suggesting that SMNFL may have specific targets in motor neurons. Moreover, these data indicate that the vulnerability of motor neurons cannot simply be ascribed to the differential expression or a more dramatic reduction of SMNFL in spinal cord when compared with brain tissue. Improving the stability of SMN7 protein might be envisaged as a new therapeutic strategy in SMA.
--------------------------------------------------------------------------------
Spinal muscular atrophies (SMAs) (Online Mendelian Inheritance of Man nos. 271150, 253550, 253400, 253300; http://www.ncbi.nlm.nih.gov/omim/) are characterized by the degeneration of lower motor neurons, leading to progressive limb and trunk paralysis associated with muscular atrophy. SMA is a frequent recessive autosomal disorder caused by mutations of the survival of motor neuron gene (SMN1).1 SMN1 is duplicated as a highly homologous gene, called SMN2, and both genes are transcribed.1 The SMN2 gene is present in all patients but is not able to compensate for SMN1 gene defects. At the genomic level, the gene dosage effect found in type I SMA, but not in type III SMA, has suggested that type I is caused by deletion of SMN1, whereas type III is associated with a conversion event of SMN1 into SMN2, leading to an increased number of SMN2 genes.1,2 This is in agreement with the tight inverted correlation between the amount of protein encoded by the SMN2 gene and the clinical severity of human SMA disease.3,4
Five nucleotides distinguish SMN2 from SMN1 without altering the amino acid sequence.1 The critical difference between these two genes is a cytosine (C) to thymine (T) transition in exon 7 of SMN2, which is responsible for the alternative splicing of exon 7 of SMN2 transcripts.5 Full-length transcripts are almost exclusively produced by the SMN1 gene, whereas the predominant form encoded by SMN2 lacks exon 7 (SMN7).1,5 Full-length transcript (SMNFL) is also encoded by SMN2 and translated into functional protein. However, it is a minor form, much less abundant than the full-length SMN1 gene product. The truncated transcript lacking exon 7 encodes a putative shorter protein in which the last 16 residues of SMNFL are replaced by four residues (EMLA) encoded by exon 8 (SMN7). Using expression vectors, in vitro experiments demonstrated that SMN7 oligomerizes less efficiently than the full-length form and that overexpressed SMN7 was unstable in a nonneuronal immortalized cell line.6,7 However, in these studies, the stability of SMN7 in humans was not elucidated.
SMN is a ubiquitously expressed protein of 294 amino acids, with a molecular mass of 38 kd. SMN forms a large multiprotein complex of 1 Md, both in the cytoplasm and in the nucleus, where it is concentrated in a structure called gem (for gemini of coiled bodies).8,9 The identification of SMN-interacting proteins of known function in nonneuronal cell lines strongly supports the view that SMN is involved in, and facilitates, cytoplasmic assembly of snRNP into the spliceosome, a large RNA-protein complex that catalyzes the splicing reaction.10,11 In the nucleus, SMN appears to be directly involved in pre-mRNA splicing, transcription, and metabolism of ribosomal RNA.11,12 More recently, it has been suggested that SMN might have an additional function in neurons related to RNA trafficking. SMN binds heterogeneous nuclear ribonucleoprotein-R (hnRNP-R), an mRNA-binding protein that may associate with ß-actin mRNA in vitro, suggesting a role for SMN in the assembly and/or transport of ß-actin mRNP complexes into growth cones.13 Moreover, SMN was localized in granules and was transported down axons of cultured neurons.14 Together, these data suggested that SMN plays a role in the metabolism of RNA in the nucleus and/or in the transport of some transcripts in axons.
Several hypotheses have been raised to explain the higher vulnerability of motor neurons in response to the SMN1 mutation. In vivo, the link between RNA metabolism and SMA pathogenesis has been suggested. However, one report did not find a single abnormal splicing pattern in mice carrying a heterozygous deletion of the Smn gene.15 Expression profiles of 8400 genes in mouse skeletal muscle and spinal cord expressing SMN7 RNA only revealed an early, and specific, up-regulation of genes involved in pre-mRNA splicing, ribosomal RNA processing, and RNA decay.16 The observed changes could represent an adaptive response of the RNA-processing machinery because of the lack of a component normally involved in the process. These results thus provide indirect evidence for a role of SMN in RNA metabolism in vivo. However, overexpression of orthologous RNA metabolism-related genes, specific to the SMN defect, was not detected in skeletal muscle or spinal cord of human SMA samples, at either fetal or postnatal stages.16 The difference between mouse and human data could be attributable to the reduction, but not the absence, of SMNFL in human SMA tissues or to a defect in the metabolism of some RNA subclasses involved in the structural specificities of motor neurons. In agreement with this hypothesis, a correlation was found between the capacity for snRNP assembly and SMN protein level in SMA type I patients.17 The link between RNA metabolism and SMA pathogenesis has been also strongly suggested by the ability of injected U snRNPs to prevent motor neuron degeneration in SMN-deficient zebrafish embryos.18 Surprisingly, more recent data have shown that the function of SMN in motor axon outgrowth of zebrafish might be dissociated from snRNP function, suggesting that the motor axon defect observed in zebrafish mutated for SMN is unlinked to snRNP biosynthesis.19 The developmental defect of motor axon growth has, however, not been reported to date in human SMA autopsy materials.
In vitro, overexpression of SMN7 seems to be proapoptotic.20,21 In addition, in chick forebrain neurons, SMN7 was unable to be transported, and overexpression of SMN7 resulted in shorter axons associated with the retention of SMN7 in the nucleus.14 These data suggested that, in addition to the deleterious effect of a reduced dose of SMNFL, SMN7 might have a dominant-negative effect on neurons.14,20,21 However, these data were not confirmed using a different culture system.22
Motor axons can reach more than 1 m long, making motor neurons one of the largest cells in the body in terms of both volume and surface area. Such large cellular units require molecular machinery able to produce and regulate molecules, including proteins and RNA, from the cell body to the neuromuscular junction, through motor axons. Neurons express high numbers of genes with respect to other cell types, and alternative splicing as well as the stability of mRNAs are important posttranscriptional mechanisms necessary to create and regulate protein diversity.23
In summary, several hypotheses have been raised to explain the vulnerability of motor neurons in SMA: i) a dominant effect of SMN7 in motor neurons; ii) a more dramatic reduction of SMN in spinal cord compared with other tissues or a higher demand of SMN in motor neurons, attributable to their large size; iii) a role for SMN in the metabolism of a subclass of RNA specific to motor neurons; or iv) an additional function of SMN in these cells, unrelated to snRNP biogenesis. Here, we generated antibodies specific to SMN7 and SMNFL to clearly determine the presence or absence of the SMN proteins in normal and disease states. We show that the SMN7 protein is undetectable in human lymphoblastoid cell lines or tissues, including neuronal tissues. In addition, we show that the stability of the SMN7 protein depends on the cell type. The SMN protein detected in SMA samples is translated only from the full-length SMN2 transcript, which results in the dramatic decreased level of SMN in all of the early fetal SMA tissues examined. In addition to therapeutic strategies aimed at either up-regulating SMN2 gene expression or at limiting exon 7 skipping, improving the stability of the SMN7 protein could represent a new attractive therapeutic strategy in SMA.

【关键词】  characterization expression stability products

Materials and Methods

Plasmids

The human SMNFL and SMN7 cDNA were amplified by polymerase chain reaction (PCR) and cloned in-frame to the 3' end of GFP, at the MluI and XmaI restriction sites of the pCX-LGN plasmid (called GFP plasmid in the text), derived from pCX-EGFP (gift from IGBMC, Strasbourg, France). The GFP, GFP-SMNFL, and GFP-SMN7 plasmids were purified using the EndoFree plasmid maxi kit (Qiagen S.A., Courtaboeuf, France), and both plasmids were checked by sequencing both strands.

Generation of Antibodies

Rabbit SMNFL- or SMN7-specific antibodies were generated against two synthetic peptides chosen in the human amino acid sequence encoded by exon 7 (peptide hsmnEx7, GFRQNQKEGRCSHSLN) or exon 6 fused to 8 (peptide hsmnEx8, GYYMEMLA), respectively. These peptides were conjugated to KLH and injected into rabbits, and antisera hSMNex7-5381 and hSMNex8-5699 were collected and purified on an affinity column (Invitrogen, Paisley, UK).

Human Cell Lines and Tissues

Lymphoblastoid Cell Lines

Three control cell lines harboring two SMN1 genes and 0 (10-253), one (10-252), or two (11-771) SMN2 genes, three type I SMA patients (10-235, 10-237, and 10-238) with two SMN2 genes, and two type III SMA patients (10-240 and 10-241) with four SMN2 genes were selected for Western blot analysis. SMA patients were born to consanguineous parents. Cells were maintained in RPMI 1640 medium, supplemented with 20% fetal bovine serum, 1% glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin (Invitrogen).

Human Fibroblasts

Primary cultures of fibroblasts were established from skin biopsies of two type I, two type II, and two type III SMA patients. Type I patients had two SMN2 genes (SMAI-1 and SMAI-2), type II patients had three SMN2 genes (SMAII-1 and SMAII-2), a type III patient (SMAIII-1) had three SMN2 genes, and the other one (SMAIII-2) had four SMN2 genes. Control fibroblasts had two SMN1 genes and two SMN2 genes (Cont-1, Cont-2). Fibroblasts were analyzed below 10 in vitro passages and maintained in Dulbecco??s modified Eagle??s medium with 4.5% glucose, supplemented with 20% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin (Invitrogen) as previously described.24

Fetal Tissues

Two fetuses predicted to have type I SMA (SMA-I-3 and SMA-I-4) and one type II SMA (SMA-II-3), as well as three fetuses affected with diseases unrelated to SMA (Cont 3 to 5), were studied. The fetuses were at 11 to 18 weeks of gestation. Samples were collected in the framework of prenatal diagnosis, following the informed consent of parents in Necker-Enfants Malades (Paris, France) or Antoine Beclre (Clamart, France) hospitals. The number of SMN1 and SMN2 genes was determined as previously described.25 Fetal tissues used for controls and SMA fetuses were distinct from those used for lymphoblastoid cell lines or for fibroblast analyses. SMA patients and fetuses carried a homozygous deletion of SMN1 exon 7 (data not shown). Informed consent was obtained from all patients and controls or from their parents.

Cell Cultures and Transfection Experiments

HeLa Cell Transfections

HeLa cells were maintained at 37??C in Dulbecco??s modified Eagle??s medium (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen), 100 U/ml penicillin, and 100 µg/ml streptomycin (Invitrogen). Cells were transfected using 1 µg of DNA (GFP, GFP-SMNFL, or GFP-SMN7 plasmids) and 2 µl of Lipofectamine 2000 (Invitrogen). Cells were harvested 24 hours after transfection.

Primary Cultures of Mouse Cortical Neurons

Primary cultures of cortical neurons were prepared from 14.5-day postcoitum C57/BL6 wild-type mouse embryos. The cortices of embryos were dissected in Mg2+, Ca2+-free Hanks?? balanced salt solution, trypsinized, dissociated, and then transfected using a Nucleofector apparatus (Amaxa Biosystems, Cologne, Germany) with 2 µg of DNA (GFP, GFP-SMNFL, or GFP-SMN7 plasmids) for 5 million cells. Transfected neurons were plated on poly-D-ornithine (5 mg/ml)- and laminin (2 mg/ml)-coated dishes, at a density of 9.3 x 104 cells/cm2 in Neurobasal medium (Invitrogen) plus B27 supplement. Cells were maintained for 6 days at 37??C in 5% CO2. Three independent transfection experiments were performed.

One, three, or six days after transfection, the same fields were analyzed by phase contrast or fluorescence using a Zeiss Axiovert 200M fluorescence microscope (Carl Zeiss, Thornwood, NY). Cell survival was estimated from the mean number of nuclei on five same phase contrast fields 1, 3, and 6 days after transfection. The total neurite length, relative to the number of nuclei per field, was analyzed 1, 3, and 6 days after plating from the same five fields by using the stereological method, as previously described.26

Cultures were fixed after 1 day with 4% paraformaldehyde for 30 minutes at 4??C, permeabilized with 0.5% Triton X-100, incubated with 5% nonfat dry milk and 0.05% Tween 20 for 1 hour at room temperature, and then incubated in anti-20S proteasome antibody (1:500; Biomol International, Exeter, UK) overnight at 4??C. After rinsing, the cultures were incubated with anti-rabbit IgG conjugated to Alexa 594 (1:500; Molecular Probes, Eugene, OR) for 1 hour at room temperature. Cultures were then mounted with Vectashield-DAPI (Vector Laboratories, Burlingame, CA) and observed under a Zeiss Axioplan 2 fluorescence microscope.

Mouse Embryonic Fibroblasts

Mouse embryonic fibroblasts were prepared from 14.5-day postcoitum C57/BL6 wild-type embryos following standard procedures and cultured in Dulbecco??s modified Eagle??s medium and 10% fetal bovine serum. Mouse embryonic fibroblasts were transfected using Nucleofector apparatus (Amaxa Biosystems) with 5 µg of DNA (GFP, GFP-SMNFL, or GFP-SMN7 plasmids) for 12.5 million cells.

Protein Stability Assay

Primary cultures of mouse cortical neurons or embryonic fibroblasts were transfected with GFP-SMNFL or GFP-SMN7 as described above. Twenty-four hours after transfection, cells were treated with 10 µg/ml cycloheximide (Sigma, St. Louis, MO) to inhibit protein synthesis. The cycloheximide-treated cells were harvested at different time points (0, 1, 3, and 6 hours after treatment) and processed for immunoblotting with anti-GFP. Anti-tubulin antibodies were used as internal control for loading.

Immunoblotting Experiments

Total proteins were extracted from HeLa cells or primary cultures of mouse cortical neurons in a buffer containing 25 mmol/L sodium phosphate, pH 7.2, 5 mmol/L ethylenediaminetetraacetic acid (EDTA), and 1% sodium dodecyl sulfate supplemented with protease inhibitor cocktail (Roche, Mannheim, Germany) and boiled for 10 minutes. For lymphoblastoid cells, proteins were extracted using 100 mmol/L Tris-HCl, pH 7.5, 1% Triton X-100, 100 mmol/L NaCl, 5 mmol/L EDTA, 0.5 mmol/L MgCl2, 1 mmol/L ß-mercaptoethanol, and 1 mmol/L phenylmethyl sulfonyl fluoride supplemented with protease inhibitor cocktail (Roche). Human fibroblast cells were resuspended in a buffer containing 62.5 mmol/L Tris-HCl, pH 6.8, 5 mmol/L EDTA, and 10% sodium dodecyl sulfate, supplemented with protease inhibitor cocktail (Roche).

Samples were passed through a 1-ml syringe with a 27-gauge needle at least 20 to 25 times. Tissues from fetuses were rapidly frozen and crushed in liquid nitrogen using a mortar and pestle. Pulverized tissue samples were transferred into 5 vol of a buffer containing 25 mmol/L sodium phosphate (pH 7.2), 5 mmol/L EDTA, and 1% sodium dodecyl sulfate supplemented with protease inhibitor cocktail (Roche), and boiled for 5 minutes. Protein samples were mixed with an equal volume of 2x Laemmli buffer, electrophoresed, and then transferred.

For SMN, 20S proteasome, and tubulin immunodetection, anti-SMNtl (1:2000 to 1:10,000; BD Transduction Laboratories, San Jose, CA), anti-20S proteasome (1:10,000; Biomol International), and anti-tubulin monoclonal antibodies (1:20,000; Sigma) were used. After washes in phosphate-buffered saline and 0.05% Tween 20, membranes were incubated with anti-mouse or anti-rabbit IgG conjugated to horseradish peroxidase (1:5000 to 1:20,000; Jackson ImmunoResearch, West Grove, PA), and the immune complexes were revealed using chemiluminescent detection reagents (Pierce, Rockford, IL). For GFP, anti-GFP polyclonal antibody (1:2000 to 1:8000; Molecular Probes) was used and revealed using anti-rabbit IgG conjugated to horseradish peroxidase (1:5000; Jackson ImmunoResearch). For the detection of SMNFL and SMN7 proteins, hSMNex7-5381 and hSMNex8-5699 purified antisera generated in this study were diluted at 1:100 and revealed by anti-rabbit IgG conjugated to horseradish peroxidase (1:20,000; Jackson ImmunoResearch).

DNA and RNA Analysis

Total DNA was extracted from primary culture of cortical neurons transfected with plasmids encoding GFP, GFP-SMNFL, or GFP-SMN7 and amplified with primers R-GFP (5'-TCTTCAAGTCCGCCATGCCC-3') and F-GFP (5'-TGTACTCCAGCTTGTGCCCC-3'). Interleukin 2 was co-amplified as an internal control using primers IMR0042 (5'-CTAGGCCACAGAATTGAAAGATCT-3') and IMR0043 (5'-GTAGTTGGAAATTCTAGCATCATCC-3').

Total RNA was extracted from lymphoblastoid cells or primary culture of cortical neurons using the TRIzol procedure (Invitrogen). First-strand cDNA synthesis was performed by using oligo(dT) primers. Negative controls without reverse transcriptase (RT) were also performed. PCR amplification analysis of single-strand cDNA was performed using primers flanking SMN exon 7 (541C544, 5'-GAGAACTCCAGGTCTCCTGGA-3' and 541Ci993, 5'-CGCTTCACATTCCAGATCTGT-3') and amplified SMN transcripts containing (450 bp) or lacking exon 7 (396 bp). Transcript analysis of cortical neurons transfected by plasmids encoding GFP, GFP-SMNFL, or GFP-SMN7 was performed using primers R-GFP (5'-TCTTCAAGTCCGCCATGCCC-3') and F-GFP (5'-TGTACTCCAGCTTGTGCCCC-3'). As internal control for RT-PCR amplification analysis, Aldolase A or ß-actin cDNA were co-amplified using primers Aldo 1 (5'-TAAGAAGGATG- GAGCCGACTTTG-3') and Aldo 600 (5'-GCGAGGCTGTTGGCCAGGGCGCG-3') or ßAHU1 (5'-CCAACCGCGAGAAGATGACCCAG-3') and ßAHU2 (5'-GGAAGAGTGCCTCAGGGCAGCG-3'), respectively. RT-PCR products were separated by agarose gel electrophoresis and labeled with ethidium bromide.

Results

Generation of New Polyclonal Antibodies Specific to SMNFL and SMN7

Peptides hSMNex7 and hSMNex8 were chosen from the human amino acid sequence encoded by SMN exon 7 or exon 6 fused to exon 8, which are specific to the SMNFL or SMN7 isoforms, respectively. The peptides were synthesized, coupled with KLH and injected into rabbits. The best immune response was obtained with antiserum hSMNex7-5381 (corresponding to peptide hSMNex7) and hSMNex8-5699 (corresponding to peptide hSMNex8) as monitored by an enzyme-linked immunosorbent assay (data not shown). Antisera were purified by immunoaffinity beads bearing the corresponding peptides. The human SMNFL and SMN7 cDNA were fused to the 3' end of GFP (called GFP-SMNFL and GFP-SMN7, respectively) and cloned into the expression plasmid pCX-LGN. Sequence analysis of cloned cDNA showed, as expected, a nucleotide sequence identical to human wild-type SMNFL or SMN7 cDNA (data not shown). Transfection experiments of HeLa cells with these plasmids, or with GFP alone, were performed. Proteins were extracted from transfected and nontransfected cells and analyzed by immunoblotting experiments using antisera hSMNex7-5381, hSMNex8-5699, or antibody against GFP (Figure 1) . hSMNex7-5381 antiserum detected both the recombinant GFP-SMNFL (65 kd) and endogenous proteins (38 kd) but not, as expected, GFP-SMN7. The hSMNex8-5699 antiserum detected the recombinant GFP-SMN7 (64 kd) but not the recombinant SMNFL proteins (Figure 1) . A faint band similar in size to that of the SMN endogenous band was inconsistently detected in HeLa cells (Figures 1, 7, and 8) . No cross-reactivity of hSMNex7-5381 and hSMNex8-5699 antisera against SMN7 and SMNFL recombinant proteins, respectively, was detected.

Figure 1. Characterization of new polyclonal antibodies directed against human SMNFL and SMN7 proteins. Immunoblot analysis of proteins extracted from HeLa cells transfected with a plasmid expressing GFP (lane 1), GFP-SMNFL (lane 2), GFP-SMN7 (lane 3), or from nontransfected HeLa cells (lane 4). The membranes were incubated with hSMNex7-5381 (A), hSMNex8-5699 (B) antisera, or with GFP and tubulin antibodies (C). An asterisk indicates a nonspecific band (see comment on Figure 7 ).

Overexpression of SMNFL or SMN7 Reveals the Dramatic Instability of SMN7 Protein in Neurons but Not in Other Cell Types in Culture

Previous reports have shown that overexpression of SMN7 resulted in shorter axons associated with nuclear retention of the SMN7 protein or in cell death. These data suggested a dominant-negative effect of SMN7.14,20,21 To determine whether this effect was observed in mammalian cells, plasmids encoding GFP-SMNFL, GFP-SMN7, or GFP only were transfected into various cell types including HeLa cells, cortical neurons, or fibroblasts from mouse embryos.

In HeLa cells, transfection experiments using these plasmids, or with GFP alone, led to the overexpression of both fused proteins, as determined by immunoblotting experiments using antisera hSMNex7-5381, hSMNex8-5699, or antibody against GFP only (Figure 1) . Microscopic examination of transfected cells revealed the presence of SMNFL or SMN7 aggregates in both the cytoplasm and the nucleus (Figure 2) . To determine whether similar findings were observed in mammalian neuronal cells, cortical neurons were collected from 14.5-day postcoitum mouse embryos and transfected with GFP-SMNFL, GFP-SMN7, or GFP constructs, plated, and observed for 6 successive days. The density of neurites was evaluated in neurons 1, 3, and 6 days after cell plating, by examining the same fields in phase contrast, and did not reveal any difference whatever with the transfected construct (Figure 3D) . These data indicate that overexpression of SMNFL or SMN7 did not have a deleterious effect on neurite outgrowth of mouse neurons in culture when compared with GFP alone. Microscopic examination of neurons transfected with GFP-SMNFL revealed aggregates of SMN in both the nucleus and the cytoplasm, including the neurites (Figure 2 and see Supplemental Figure 1 at http://ajp.amjpathol.org). Surprisingly, however, in cortical neurons transfected with GFP-SMN7, no or very weak fluorescent signal was observed even after immunolabeling of GFP (Figure 2 and see Supplemental Figure 2 at http://ajp.amjpathol.org; and data not shown).

Figure 2. Expression and subcellular localization of SMNFL and SMN7 in various mammalian cell types. Primary cultures of mouse cortical neurons (ACC'), mouse fibroblasts (DCF'), and HeLa cells (GCI') were transfected with a plasmid expressing GFP-SMNFL (A, D, and G), GFP-SMN7 (B, E, and H), or GFP (C, F, and I). GFP-SMNFL is concentrated as aggregates in the cytoplasm and in the nucleus of neurons (A), fibroblasts (D), and HeLa cells (G). GFP-SMN7 is localized in both the nucleus and cytoplasm of fibroblasts (E) and HeLa cells (H). B: Despite normal growth and survival of GFP-SMN7-transfected neurons, only a very weak signal can be seen in the cytoplasm. Nuclei are stained with 4,6-diamidino-2-phenylindole. A'CC', D'CF', and G'CI' are merged images. Scale bars: 10 µm (ACC', GCI'); 20 µm (DCF').

Figure 3. Characterization of primary cultures of cortical neurons transfected with a plasmid expressing GFP-SMNFL (lane 1), GFP-SMN7 (lane 2), and nontransfected control (lane 3). Quantification of plasmid DNA through PCR amplification of GFP sequence (A), RNA using RT-PCR amplification (B), or GFP fusion protein by immunoblot experiment (C) 1 day after plating. Although a similar amount of transfected plasmid was used, a marked decrease in RNA (48%) and protein (81%) encoding GFP-SMN7 was observed when compared with GFP-SMNFL. Relative amounts of plasmid DNA, RNA encoding GFP, and GFP protein were evaluated by determining the ratio of transgene to internal control (interleukin for DNA, aldolase for RNA, and tubulin for protein analyses). For transcript analysis, PCR amplification was performed with (+) or without (C) reverse transcriptase as a negative control. D: Cell viability was estimated from the mean number of nuclei on five same-phase contrast fields at 1, 3, and 6 days of culture. The neurite outgrowth was measured 1, 3, and 6 days after plating of transfected cortical neurons by a stereological procedure on five same-phase contrast fields. No statistically significant difference was observed whatever the transfected plasmids or the day of examination. Three independent experiments were performed. Bar means SE.

To clarify the reasons for such a dramatic reduction in GFP-SMN7 protein in neurons, semiquantitative analysis of DNA, RNA, and protein encoding GFP-SMNFL or GFP-SMN7 was performed 1 day after the transfection of cortical neurons (Figure 3) . Despite similar levels of plasmid DNA, reduced levels of GFP-SMN7 RNA (48%) and protein (81% reduction, P = 0.001) were detected when compared with the GFP-SMNFL RNA or protein levels and internal controls (Figure 3) . To determine whether these data might result from cell death, the survival of neurons transfected with GFP-SMN7, GFP-SMNFL, or GFP alone was analyzed. The survival of neurons was similar regardless of the plasmid used or the day of examination (from days 1 to 6, Figure 3 ). These data indicate a marked instability of GFP-SMN7 RNA and protein in neurons, but not in HeLa cells, when compared with either GFP-SMNFL or GFP (Figures 1 and 3) .

To determine whether this instability is found in other cell types, the same constructs were transfected into a primary culture of fibroblasts from wild-type mouse embryos (Figure 2) . Interestingly, fibroblasts transfected with GFP-SMNFL display numerous aggregates mainly in the cytoplasm, with very few in the nucleus. In contrast, numerous aggregates were detected in both the nucleus and the cytoplasm of fibroblasts transfected with the GFP-SMN7 construct. These data indicate that the stability of SMN7 protein depends on the cell type; it is much more unstable in neurons than in the other cell types.

To compare further the protein stability of GFP-SMNFL and GFP-SMN7 proteins in different cell types, primary cultures of mouse cortical neurons and fibroblasts were transfected with GFP-SMNFL and GFP-SMN7 plasmids and then treated with cycloheximide to inhibit protein synthesis (Figure 4) . The levels of fused proteins have been analyzed 0, 1, 3, and 6 hours after treatment and compared with that of tubulin used as internal control. A dramatic reduction of the relative amount of GFP-SMN7 was observed in neurons when compared with the full-length protein (GFP-SMNFL). Indeed, the GFP-SMN7 protein disappeared almost completely 3 hours after treatment contrasting with the stability of the GFP-SMNFL (Figure 4) . In fibroblasts, however, the stability of GFP-SMN7 and GFP-SMNFL was similar (Figure 4) . These data indicate that the stability of the SMN7 protein is regulated in a cell-dependent manner.

Figure 4. SMN protein stability assay in neurons and fibroblasts. Primary cultures of mouse cortical neurons (A) and embryonic fibroblasts (B) were transfected with a plasmid expressing GFP-SMNFL or GFP-SMN7. Cells were treated with cycloheximide (CHX) 24 hours after transfection, and proteins were harvested at indicated time points (0, 1, 3, and 6 hours after CHX treatment). Similar amounts of total protein extracts were loaded and analyzed by Western blot using anti-GFP and anti-tubulin antibodies. C: Levels of GFP-SMNFL, GFP-SMN7, and tubulin proteins were evaluated by densitometric analysis. Each point corresponds to the ratio of the fused protein: tubulin relative to that at time point 0.

To determine whether the instability of the SMN7 protein is linked to the activation of the ubiquitin proteasome pathway, immunolabeling and Western blot analysis of the 20S proteasome was performed in HeLa and cortical neurons transfected with GFP-SMN7or GFP-SMNFL. No change in the expression or subcellular localization of the 20S proteasome was observed under any of the conditions analyzed (Figure 5 and see Supplemental Figure 3 at http://ajp.amjpathol.org).

Figure 5. Immunolabeling experiment of the 20S proteasome in HeLa and cortical neurons. Primary cultures of mouse cortical neurons (ACB''') or HeLa cells (CCD''') were transfected with a plasmid expressing GFP-SMNFL (ACA''' and CCC''') or GFP-SMN7 (BCB''' and DCD'''). The 20S proteasome is labeled in red (A'CD') and GFP fluorescence is shown in green (ACD). Nuclei were stained with 4,6-diamidino-2-phenylindole (A'''CD'''). A'' to D'' are merged images. Scale bars: 10 µm (A-B'''); 20 µm (CCD''').

Tight Correlation between SMNFL and SMN2 Gene Number and the Lack of SMN7 Protein in Human SMA Cell Lines

Previous studies have shown a correlation between SMN protein level and the severity of the disease.3,4,24 However, without specific antibodies to SMNFL or SMN7, it was not possible to determine whether the detected protein corresponded to SMNFL, SMN7, or both protein isoforms in SMA patients. In this study, lymphoblastoid cell lines from controls carrying two SMN1 genes and none to two SMN2 genes, and SMA patients lacking SMN1 gene but carrying two or four SMN2 genes, were analyzed. SMA patients were born to consanguineous parents and shared common haplotypes at the 5q13 region, indicating that the genomic organization at the SMN2 locus was identical on both 5q13 alleles.1 RT-PCR analysis of SMN transcripts revealed, as expected, the absence of SMN7 transcripts in controls without the SMN2 gene, and a higher level of SMN7 in SMA patients (Figure 6) . Interestingly, similar levels of full-length and truncated SMN RNAs were observed in controls lacking SMN2 but carrying two SMN1 genes and in SMA patients lacking SMN1 but carrying two SMN2 genes (Figure 6 and data not shown), respectively. These data indicate that in vivo SMNFL and SMN7 RNAs have similar stability.

Figure 6. Correlation of SMNFL protein level with SMN2 gene copy number and the absence of SMN7 protein in lymphoblastoid cell lines. Immunoblot analysis of proteins extracted from lymphoblastoid cell lines derived from controls, type I, or type III SMA patients. Immunoblots were incubated with hSMNex7-5381 (A), SMNtl monoclonal antibody (B), or with hSMNex8-5699 (C). Tubulin was used as internal control for loading. D: RT-PCR analysis of RNA extracted from the same cell lines, using primers flanking SMN exon 7, showing the absence of SMN7 transcripts in the 10-253 cell line and the reduction of SMNFL transcripts in SMA patients. ß-Actin was used as internal control. The 10-253 cell line, lacking the SMN2 gene, was used as a negative control for the detection of SMN7 RNA or protein. E: The number of SMN1 or SMN2 genes in each control or patient is shown.

Proteins were extracted from cell lines, and immunoblotting experiments were performed using antibodies directed against the N-terminal part of the SMN protein (SMNtl), against exon 7 (hSMNex7-5381), or against SMN7 (hSMNex8-5699; Figure 6 ). Importantly, our data did not reveal any signal at the expected size using antiserum against SMN7 protein. This indicates that the SMN protein detected by SMNtl or hSMNex7-5381 antisera corresponds to the SMNFL form only. Interestingly, this analysis shows a tight correlation between SMNFL protein level and the number of SMN2 genes in both controls and SMA patients (Figure 6) . An increased amount of SMNFL was observed in controls carrying an identical number of SMN1 genes (n = 2) but an increased number of SMN2 genes (from 0 to 2). A similar correlation was observed in SMA patients. These data demonstrate the absence of SMN7 in lymphoblastoid cell lines from controls and SMA patients. The data also reveal a contribution of the SMNFL protein, encoded by the SMN2 gene, to the total amount of SMN in both controls and SMA patients.

To determine whether the protein profiles in other cell types were similar, immunoblotting experiments of proteins extracted from fibroblasts of controls and SMA born to unrelated parents (Figure 7) were performed. No SMN7 was detectable either in these cell lines. Although a marked reduction of SMNFL protein was observed in the SMA type I patients, type II or type III patients, carrying the same number of SMN2 genes, may express very low or high level of SMNFL (Figure 7) .

Figure 7. Absence of SMN7 protein in fibroblasts of controls and SMA patients. Immunoblot analysis of total protein extracts from fibroblasts derived from controls or type I, II, or III SMA patients. A: The membranes were incubated with hSMNex8-5699 (top), SMNtl (middle), or tubulin (bottom). The 10-253 lymphoblastoid cell line was used as a negative control for the SMN7 protein. HeLa cells transfected with the plasmid expressing GFP-SMN7 (lane 2) were used as a positive control for the detection of SMN7 protein when compared with GFP only (lane 1). B:SMN1 and SMN2 gene copy number in each control or patient. Asterisk indicates a nonspecific band because the 10-253 individual does not carry the SMN2 gene from which SMN7 transcript is produced.

Absence of SMN7 Protein and Dramatic Reduction of SMNFL in All Human SMA Tissues Examined

To determine whether the SMN7 protein exists in human tissues, Western blot analysis was performed on proteins extracted from various tissues from aborted fetuses predicted to have type I SMA or a disease unrelated to SMA (Figure 8) . Several tissues were analyzed from both SMA and control, including skeletal muscle, liver, brain, and spinal cord. No SMN7 was detectable using antibody specific to SMN7 (Figure 8) . These data indicate that SMN7 is absent, or present at undetectable levels, in SMA tissues.

Figure 8. Absence of SMN7 protein and dramatic reduction of SMNFL protein levels in SMA fetuses. Immunoblot analysis of proteins extracted from spinal cord (A) or liver (B). The membranes were incubated with hSMNex8-5699 antiserum specific to human SMN7 protein. The 10-253 lymphoblastoid cell line (Lymph.) was used as a negative control and HeLa cells transfected with a plasmid expressing GFP-SMN7 (lane 2) as a positive control when compared with GFP only (lane 1). C: Immunoblot analysis of proteins extracted from various tissues of the same type I SMA fetus compared with those of control fetus. Asterisk indicates a nonspecific band (see comment on Figure 7 ).

In addition, a dramatic reduction in SMNFL was observed in SMA samples, regardless of the tissue type examined (Figure 8C) . Interestingly, Western blot analysis of SMN did not reveal any difference in the level of SMNFL between brain and spinal cord. Thus the vulnerability of the spinal cord in SMA cannot be ascribed to the differential expression or a more dramatic reduction of SMNFL in spinal cord when compared with brain tissue. Moreover, our data show a dramatic reduction in the level of SMN protein early during development.

Discussion

Here, we show that the SMN7 protein is absent or present at very low level in human controls, in SMA tissues, and in cell lines. These results exclude the hypothesis of a dominant effect of SMN7 to explain the vulnerability of motor neurons in SMA. These data are, however, consistent with the absence of symptoms in individuals carrying one or two SMN1 genes and a varying number of SMN2 genes. In addition, these data are in agreement with the correlation between the number of SMN2 genes and the clinical expression of the disease; a high number of SMN2 genes being associated with the mildest form of SMA.1,2,27

The lack of SMN7 protein is attributable to the instability of SMN7 at both RNA and protein levels, depending on the cell type. The instability of SMN7 RNA was strongly suggested after semiquantitative analysis of SMN7 RNA of murine neurons transfected with GFP-SMN7. A 48% reduction in SMN7 RNA level was observed when compared with SMNFL RNA levels. In addition to RNA instability, SMN7 protein is unstable when compared with SMNFL in transfected neurons in culture (81% reduction). In contrast, no difference was detected between SMNFL RNA levels in control human lymphoblastoid cell lines carrying two SMN1 genes but lacking SMN2, and the level of SMN7 RNA in SMA patients carrying two SMN2 genes but no SMN1. These data confirm that human SMN7 and SMNFL RNA have similar stability (P.S.V., unpublished data).28

Our study shows that the stability of the SMN7 protein depends on the cell type. In transfection experiments, SMN7 protein was detected in HeLa cells and in primary cultures of mouse fibroblasts whereas only very low levels of protein were detectable in primary cultures of mouse neurons. Moreover, the protein stability assay clearly showed a dramatic instability of the GFP-SMN7 protein in neurons but not in fibroblasts. These data indicate that the stability of SMN7 is regulated in a cell-dependent manner. The marked instability of SMN7 protein in neurons but not in other cell types could explain the higher vulnerability of motor neurons in SMA in the case of a putative residual SMN7 function, although cortical but not spinal cord motor neurons were examined. However, in vivo, SMN7 protein was not detected in either human neuronal (including spinal cord) or nonneuronal tissues. These data indicate that motor neuron degeneration is caused uniquely by a reduction in SMNFL protein levels.

An important question is whether improving SMN7 protein stability could have a beneficial effect in the treatment of SMA. In vivo, homozygous deletions of murine Smn exon 7 in all cell types results in embryonic lethality at 9 days postcoitum, whereas Smn knockout resulted in earlier embryonic lethality (3 days postcoitum) suggesting that, in vivo, SMN7 protein might have a residual function.29,30 These data are further supported by the improved median survival of (SmnC/C, SMN2) mutant mice overexpressing SMN7.31 However, SMN7 is unable to compensate for the complete absence of SMNFL as shown by the severe phenotype of mice carrying a homozygous deletion of Smn exon 7 in various cell types.29,32,33 These data indicate that overexpression of SMN7 could have a beneficial effect on SMA. In addition to therapeutic strategies aimed at up-regulating SMN2 gene expression or limiting exon 7 skipping,34 our study allowed the identification of a new target consisting in improving the stability of SMN7 protein in SMA patients. These new polyclonal antibodies specific to the SMN7 protein should be very useful for the detection of drugs able to stabilize this isoform.

It has been hypothesized that the vulnerability of motor neurons in SMA might be related to the differential level of expression of the SMN protein in the spinal cord compared with other tissues. SMN protein analyses in various human controls and SMA fetal tissues presented here do not support this hypothesis. The SMN protein level in SMA fetuses is dramatically reduced at an early stage of spinal cord development, whereas symptoms occur postnatally even in the most severe forms of the disease. These data suggest that during development a protein involved in the same molecular pathway can compensate for the dramatic reduction in SMN. Alternatively, motor neurons may require more SMN protein for axonal outgrowth or maintenance during the postnatal period, which would be in agreement with a link between SMN and axonal transport.

A tight correlation between the amount of SMN protein and the clinical severity of SMA was clearly demonstrated.3,4 In addition, clinical severity is correlated with the level of SMNFL RNA and the number of SMN2 genes (P.S.V, unpublished data).35 Our study shows a tight correlation between the number of SMN2 genes (from 0 to 4) and SMNFL protein level in both controls and in SMA patients born to consanguineous parents. In these patients, the organization of the SMN locus, including the position of the SMN2 genes on each chromosome, is identical on both alleles. In contrast, in patients born to nonconsanguineous parents, the SMN2 gene may have a telomeric, centromeric, or an ectopic position on one allele and a different configuration on the other allele. This could explain the weak correlation in SMA patients born to nonconsanguineous parents. Although the sequence of the promoter of SMN1 and SMN2 genes is identical,36,37 cis-regulatory elements could be located far from the gene promoter, which, in addition to possible transacting factors, could regulate expression either at the telomeric locus or at the centromeric locus. The promoter activity of the SMN2 gene could therefore depend on its genomic environment and thus modulate the response of drugs acting at the level of the gene promoter. A refined analysis of the SMN2 gene by mapping the 5q13 locus in SMA patients should resolve this issue.

Despite the generation and characterization of several model systems, SMA pathogenesis remains unclear. This is probably attributable to the difficulty in generating model systems that can mimic the situation of human SMA, namely the presence of the SMN2 locus in its genomic and cellular contexts (motor neurons). Here, we exclude two hypotheses that were proposed from in vitro data or transfected cell systems. The generation of novel cellular models, derived from human pluripotent cells able to differentiate into motor neurons, will be very helpful in determining whether motor neuron degeneration in SMA is linked to a defect in metabolism of a subclass of RNA or an additional defective function of SMN unrelated to snRNP biogenesis.

Acknowledgements

We thank B. Desforges, V. Joshi, T. Larmonier, and V. Vongsa for assistance; and C. Caloustian for sequencing facility.

【参考文献】
  Lefebvre S, Burglen L, Reboullet S, Clermont O, Burlet P, Viollet L, Benichou B, Cruaud C, Millasseau P, Zeviani M, Le Paslier D, Frezal J, Cohen D, Weissenbach J, Munnich A, Melki J: Identification and characterization of a spinal muscular atrophy-determining gene. Cell 1995, 80:155-165

Campbell L, Potter A, Ignatius J, Dubowitz V, Davies K: Genomic variation and gene conversion in spinal muscular atrophy: implications for disease process and clinical phenotype. Am J Hum Genet 1997, 61:40-50

Lefebvre S, Burlet P, Liu Q, Bertrandy S, Clermont O, Munnich A, Dreyfuss G, Melki J: Correlation between severity and SMN protein level in spinal muscular atrophy. Nat Genet 1997, 16:265-269

Coovert DD, Le TT, McAndrew PE, Strasswimmer J, Crawford TO, Mendell JR, Coulson SE, Androphy EJ, Prior TW, Burghes AH: The survival motor neuron protein in spinal muscular atrophy. Hum Mol Genet 1997, 6:1205-1214

Lorson CL, Hahnen E, Androphy EJ, Wirth B: A single nucleotide in the SMN gene regulates splicing and is responsible for spinal muscular atrophy. Proc Natl Acad Sci USA 1999, 96:6307-6311

Lorson CL, Strasswimmer J, Yao JM, Baleja JD, Hahnen E, Wirth B, Le T, Burghes AH, Androphy EJ: SMN oligomerization defect correlates with spinal muscular atrophy severity. Nat Genet 1998, 19:63-66

Lorson CL, Androphy EJ: An exonic enhancer is required for inclusion of an essential exon in the SMA-determining gene SMN. Hum Mol Genet 2000, 9:259-265

Liu Q, Dreyfuss G: A novel nuclear structure containing the survival of motor neurons protein. EMBO J 1996, 15:3555-3565

Liu Q, Fischer U, Wan F, Dreyfuss G: The spinal muscular atrophy disease gene product, SMN, and its associated protein SIP1 are in a complex with spliceosomal snRNP proteins. Cell 1997, 90:1013-1021

Fischer U, Liu Q, Dreyfuss G: The SMN-SIP1 complex has an essential role in spliceosomal snRNP biogenesis. Cell 1997, 90:1023-1029

Eggert C, Chari A, Laggerbauer B, Fischer U: Spinal muscular atrophy: the RNP connection. Trends Mol Med 2006, 12:113-121

Pellizzoni L, Kataoka N, Charroux B, Dreyfuss G: A novel function for SMN, the spinal muscular atrophy disease gene product, in pre-mRNA splicing. Cell 1998, 95:615-624

Rossoll W, Jablonka S, Andreassi C, Kroning AK, Karle K, Monani UR, Sendtner M: Smn, the spinal muscular atrophy-determining gene product, modulates axon growth and localization of ß-actin mRNA in growth cones of motoneurons. J Cell Biol 2003, 163:801-812

Zhang HL, Pan F, Hong D, Shenoy SM, Singer RH, Bassell GJ: Active transport of the survival motor neuron protein and the role of exon-7 in cytoplasmic localization. J Neurosci 2003, 23:6627-6637

Jablonka S, Schrank B, Kralewski M, Rossoll W, Sendtner M: Reduced survival motor neuron (Smn) gene dose in mice leads to motor neuron degeneration: an animal model for spinal muscular atrophy type III. Hum Mol Genet 2000, 9:341-346

Olaso R, Joshi V, Fernandez J, Roblot N, Courageot S, Bonnefont JP, Melki J: Activation of RNA metabolism-related genes in mouse but not human tissues deficient in SMN. Physiol Genomics 2006, 24:97-104

Wan L, Battle DJ, Yong J, Gubitz AK, Kolb SJ, Wang J, Dreyfuss G: The survival of motor neurons protein determines the capacity for snRNP assembly: biochemical deficiency in spinal muscular atrophy. Mol Cell Biol 2005, 25:5543-5551

Winkler C, Eggert C, Gradl D, Meister G, Giegerich M, Wedlich D, Laggerbauer B, Fischer U: Reduced U snRNP assembly causes motor axon degeneration in an animal model for spinal muscular atrophy. Genes Dev 2005, 19:2320-2330

Carrel TL, McWhorter ML, Workman E, Zhang H, Wolstencroft EC, Lorson C, Bassell GJ, Burghes AH, Beattie CE: Survival motor neuron function in motor axons is independent of functions required for small nuclear ribonucleoprotein biogenesis. J Neurosci 2006, 26:11014-11022

Vyas S, Bechade C, Riveau B, Downward J, Triller A: Involvement of survival motor neuron (SMN) protein in cell death. Hum Mol Genet 2002, 11:2751-2764

Kerr DA, Nery JP, Traystman RJ, Chau BN, Hardwick JM: Survival motor neuron protein modulates neuron-specific apoptosis. Proc Natl Acad Sci USA 2000, 97:13312-13317

Cisterni C, Kallenbach S, Jordier F, Bagnis C, Pettmann B: Death of motoneurons induced by trophic deprivation or by excitotoxicity is not prevented by overexpression of SMN. Neurobiol Dis 2001, 8:240-251

Stamm S, Zhang MQ, Marr TG, Helfman DM: A sequence compilation and comparison of exons that are alternatively spliced in neurons. Nucleic Acids Res 1994, 22:1515-1526

Andreassi C, Angelozzi C, Tiziano FD, Vitali T, De Vincenzi E, Boninsegna A, Villanova M, Bertini E, Pini A, Neri G, Brahe C: Phenylbutyrate increases SMN expression in vitro: relevance for treatment of spinal muscular atrophy. Eur J Hum Genet 2004, 12:59-65

Saugier-Veber P, Drouot N, Lefebvre S, Charbonnier F, Vial E, Munnich A, Frebourg T: Detection of heterozygous SMN1 deletions in SMA families using a simple fluorescent multiplex PCR method. J Med Genet 2001, 38:240-243

Rønn LC, Ralets I, Hartz BP, Bech M, Berezin A, Berezin V, Moller A, Bock E: A simple procedure for quantification of neurite outgrowth based on stereological principles. J Neurosci Methods 2000, 100:25-32

Kolb SJ, Gubitz AK, Olszewski RF, Jr, Ottinger E, Sumner CJ, Fischbeck KH, Dreyfuss G: A novel cell immunoassay to measure survival of motor neurons protein in blood cells. BMC Neurol 2006, 6:6

Feldkötter M, Schwarzer V, Wirth R, Wienker TF, Wirth B: Quantitative analyses of SMN1 and SMN2 based on real-time LightCycler PCR: fast and highly reliable carrier testing and prediction of severity of spinal muscular atrophy. Am J Hum Genet 2002, 70:358-368

Baron-Delage S, Abadie A, Echaniz-Laguna A, Melki J, Beretta L: Interferons and IRF-1 induce expression of the survival motor neuron (SMN) genes. Mol Med 2000, 6:957-968

Frugier T, Tiziano FD, Cifuentes-Diaz C, Miniou P, Roblot N, Dierich A, Le Meur M, Melki J: Nuclear targeting defect of SMN lacking the C-terminus in a mouse model of spinal muscular atrophy. Hum Mol Genet 2000, 9:849-858

Schrank B, Gotz R, Gunnersen JM, Ure JM, Toyka KV, Smith AG, Sendtner M: Inactivation of the survival motor neuron gene, a candidate gene for human spinal muscular atrophy, leads to massive cell death in early mouse embryos. Proc Natl Acad Sci USA 1997, 94:9920-9925

Le TT, Pham LT, Butchbach ME, Zhang HL, Monani UR, Coovert DD, Gavrilina TO, Xing L, Bassell GJ, Burghes AH: SMNDelta7, the major product of the centromeric survival motor neuron (SMN2) gene, extends survival in mice with spinal muscular atrophy and associates with full-length SMN. Hum Mol Genet 2005, 14:845-857

Cifuentes-Diaz C, Frugier T, Tiziano FD, Lacene E, Roblot N, Joshi V, Moreau MH, Melki J: Deletion of murine SMN exon 7 directed to skeletal muscle leads to severe muscular dystrophy. J Cell Biol 2001, 152:1107-1114

Vitte JM, Davoult B, Roblot N, Mayer M, Joshi V, Courageot S, Tronche F, Vadrot J, Moreau MH, Kemeny F, Melki J: Deletion of murine Smn exon 7 directed to liver leads to severe defect of liver development associated with iron overload. Am J Pathol 2004, 165:1731-1741

Frugier T, Nicole S, Cifuentes-Diaz C, Melki J: The molecular bases of spinal muscular atrophy. Curr Opin Genet Dev 2002, 12:294-298

Sumner CJ, Kolb SJ, Harmison GG, Jeffries NO, Schadt K, Finkel RS, Dreyfuss G, Fischbeck KH: SMN mRNA and protein levels in peripheral blood: biomarkers for SMA clinical trials. Neurology 2006, 66:1067-1073

Echaniz-Laguna A, Miniou P, Bartholdi D, Melki J: The promoters of the survival motor neuron gene (SMN) and its copy (SMNc) share common regulatory elements. Am J Hum Genet 1999, 64:1365-1370

Monani UR, McPherson JD, Burghes AH: Promoter analysis of the human centromeric and telomeric survival motor neuron genes (SMNC and SMNT). Biochim Biophys Acta 1999, 1445:330-336

作者单位：From the Molecular Neurogenetics Laboratory,* INSERM U798, Evry and Paris 11 Universities, Evry, France; the Department of Genetics, Rouen University Hospital and INSERM U614, Institute for Biochemical Research, University of Rouen, Rouen, France; Hôpital Necker-Enfants Malades, INSERM, U393,